Development of efficient chemistry to generate site-specific disulfide

Subscriber access provided by TUFTS UNIV

Article

Development of efficient chemistry to generate site-specific disulfide-linked protein- and peptide-payload conjugates: Application to THIOMAB™ antibody-drug conjugates Jack David Sadowsky, Thomas H Pillow, Jinhua Chen, Fang Fan, Changrong He, Yanli Wang, Gang Yan, Hui Yao, Zijin Xu, Shanique Martin, Donglu Zhang, Phillip Chu, Josefa dela Cruz-Chuh, Aimee O'Donohue, Guangmin Li, Geoffrey Del Rosario, Jintang He, Luna Liu, Carl K. Ng, Dian Su, Gail D. Lewis Phillips, Katherine Ruth Kozak, Shang-Fan Yu, Keyang Xu, Douglas Leipold, and John S. Wai Bioconjugate Chem., Just Accepted Manuscript • DOI: 10.1021/acs.bioconjchem.7b00258 • Publication Date (Web): 21 Jun 2017 Downloaded from http://pubs.acs.org on June 22, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Bioconjugate Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Bioconjugate Chemistry

Development of efficient chemistry to generate sitespecific disulfide-linked protein- and peptidepayload conjugates: Application to THIOMAB™ antibody-drug conjugates Jack D. Sadowsky,1,* Thomas H. Pillow,1 Jinhua Chen,2 Fang Fan,2 Changrong He,2 Yanli Wang,2 Gang Yan,2 Hui Yao,2 Zijin Xu,2 Shanique Martin,1,3 Donglu Zhang,1 Phillip Chu,1 Josefa dela Cruz-Chuh,1 Aimee O’Donohue,1 Guangmin Li,1 Geoffrey Del Rosario,1 Jintang He,1 Luna Liu,1 Carl Ng,1 Dian Su,1 Gail D. Lewis Phillips,1 Katherine R. Kozak,1 Shang-Fan Yu,1 Keyang Xu,1 Douglas Leipold,1 John Wai2 1Genentech,

Inc., 1 DNA Way, South San Francisco, CA 94080 AppTec Co., Ltd, 288 Fute Zhong Road, Waigaoqiao Free Trade Zone, Shanghai 200131, PR China

2WuXi

* Corresponding author: [email protected] Abstract Conjugation of small molecule payloads to specific cysteine residues on proteins via a disulfide bond represents an attractive strategy to generate redox-sensitive bioconjugates, which have value as potential diagnostic reagents or therapeutics. Advancement of such “direct-disulfide” bioconjugates to the clinic necessitates chemical methods to form disulfide connections efficiently, without byproducts. The disulfide connection must also be resistant to premature cleavage by thiols prior to arrival at the targeted tissue. We show here that commonly-employed methods to generate direct disulfide-linked bioconjugates are inadequate for addressing these challenges. We describe our efforts to optimize direct-disulfide conjugation chemistry, focusing on the generation of conjugates between cytotoxic payloads and cysteine-engineered antibodies (i.e., THIOMAB™ antibody-drug conjugates, or TDCs). This work culminates in the development of novel, highyielding conjugation chemistry for creating direct payload disulfide connections to any of several Cys mutation sites in THIOMAB™ antibodies or to Cys sites in other biomolecules (e.g., human serum albumin and cell-penetrating peptides). We conclude by demonstrating that hindered direct disulfide TDCs with two methyl groups adjacent to the disulfide, which have heretofore not been described for any bioconjugate, are more stable and more efficacious in`mouse tumor xenograft studies than less hindered analogs.

Introduction

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

In recent years, intense focus in academia and industry has been placed on using proteins or peptide carriers to deliver bioactive small molecule “payloads” to specific tissues of the human body.1-4 Applications for such bioconjugates include imaging agents, diagnostic tools and therapeutics for cancer or other diseases.5-11 Successful development of bioconjugates relies upon efficient chemistry to attach the small molecule to the protein or peptide, ideally site-specifically.12 Of paramount importance, especially for therapeutic applications, is stability of the connection between cargo and carrier in biological milieus. For antibody-drug conjugates (ADCs), which employ an antibody as targeting agent connected to a cytotoxic small molecule payload, an unstable connection between antibody and “drug” can result in reduced efficacy and/or increased toxicity.13-14 Thus, much effort in the ADC field has been expended on improving the stability of the bond between payload and antibody. Linking via a disulfide bond is a well-recognized and attractive strategy for the generation of ADCs and other bioconjugates.15-21 Ideally, the disulfide is designed to be stable in circulation (for in vivo applications) and reduced in the cytoplasm of the target cell or tissue, releasing the payload. However, given the vulnerability of disulfide bonds to nucleophilic attack by circulating thiols such as cysteine, glutathione and albumin, it has been challenging to develop stable and highly effective disulfide-linked bioconjugates for in vivo applications. Most disulfide ADCs reported employ heterobifunctional linkers joining the small molecule payload and the antibody. In these cases, the disulfide bond is positioned between a lysinereactive handle (e.g., NHS ester), which is reacted with the antibody, and the payload. The result is a heterogeneous conjugate in which several lysines on the antibody are modified. The in vivo stability of the disulfide bond in such heterobifunctional crosslinkers has been enhanced by incorporation of methyl groups on either side of the disulfide, which block attack by thiols. However, since disulfide cleavage is also the mechanism by which the drug is released in the target cell, a compromise between high stability in circulation (more methyl groups) and efficient drug release in the target cell (fewer methyl groups) is needed to achieve maximal efficacy.16 Bioconjugates in which the disulfide bond is formed directly between an engineered Cys on the protein carrier (e.g., THIOMAB™ antibody) and thiol on the small molecule offer several advantages relative to conjugates employing heterobifunctional linkers.15, 18, 22 Such “direct disulfide” conjugation enables sitespecific attachment of the payload to the carrier, resulting in a more homogeneous product and, for ADCs, may lead to better therapeutic indices. We have observed for direct disulfide THIOMAB™ antibody-drug conjugates (TDCs) that the conjugation site on the antibody can offer protection of the disulfide in circulation, acting effectively as a steric block to disulfide cleavage.18 Upon antibody internalization and degradation in the target cell, the hindrance provided by the antibody is removed allowing facile drug release. Thus, unlike for heterobifunctional linkers, stability in circulation and release of drug in the target cell are decoupled, allowing for simultaneous maximization of both stability and release. Here we report our


Page 2 of 28

Page 3 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


efforts to optimize direct disulfide conjugation of small molecule cytotoxic payloads to Cys-engineered proteins, focusing on THIOMAB™ antibody-drug conjugates. Results and discussion The most common method to generate direct disulfide-linked bioconjugates involves the substitution reaction between a Cys on the protein and an activated disulfide on the payload (Pathway A in Figure 1). Attack of the protein Cys on the payload side of the disulfide bond gives rise to the desired conjugate and a leaving group. The 2-mercaptopyridyl (PDS) group and electron-deficient variants (e.g., 5nitro-PDS) are commonly employed in such reactions, although to our knowledge no studies have evaluated or optimized bioconjugation efficiency with these or other leaving groups in depth. THIOMAB™ an4body R

LG

Pathway A

R

S

+

S

SH

HS

S S

S S

R

LG byproduct

Payload S S

Pathway B

R

S S

S S

R

+

HS

S S

Direct disulfidelinked TDC

Figure 1. Synthetic approaches to direct disulfide THIOMAB™ antibody-drug conjugates (TDCs). In Pathway A, the THIOMAB™ antibody Cys is reacted with an activated payload disulfide, formation of the leaving group (LG) byproduct can occur if the antibody Cys reacts with the LG side of the disulfide on the payload. In Pathway B, an activated disulfide form of the THIOMAB™ antibody (generated in a previous step) is reacted with a thiol payload.

Disulfide conjugation with PDS leaving groups To get a broad sense of disulfide conjugation efficiency, we attempted to generate TDCs with ten activated disulfide-linker-payloads in which the cytotoxin, linker, leaving group and degree of methyl substitution next to the disulfide (steric hindrance) were varied (compounds 1-10, Table 1). The cytotoxin was either a maytansinoid (1-5) or monomethyl auristatin E (MMAE, 6-10), both potent inhibitors of tubulin polymerization that are being actively pursued as payloads in ADC drug discovery.23 For maytansinoid disulfide-linker-payloads, steric hindrance



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

adjacent to the disulfide was varied by the presence of zero, one or two methyl groups on the linker, respectively. These linker-payloads were designed to release upon disulfide cleavage in the target cell the free thiol DM1, DM3, or DM4,


Page 4 of 28

Page 5 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 1. Activated disulfide MMAE and DM4 payloads and associated DAR values upon conjugation to anti-Her2LC-K149C.

#

R

X

Y

n

DAR

#

1

H

H

1

1.7

2

Me

H

2

3

H

H

4

Me

5

R

X

Y

n

DAR

13

Me

Me

-

0.2

1.3

14

Me

Me

-

0.5

1

1.9

15

Me

Me

-

0.4

H

2

1.9

16

Me

Me

-

0.4

Me

Me

2

0.1

17

Me

Me

-

0.5

6

H

H

-

1.7

18

Me

Me

-

1.0

7

Me

H

-

1.6

19

Me

Me

-

0.1

8

H

H

-

1.9

20

Me

Me

-

1.8

9

Me

H

-

1.9

21

Me

Me

-

1.9

10

Me

Me

-

0.1

22

Me

Me

2

1.8

11

Me

Me

-

0

23

Me

Me

2

1.8

12

Me

Me

-

0.8

Note: For compounds 2, 4 and 7 stereochemistry at the carbon indicated with an asterisk is undefined



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

A

Page 6 of 28

B DAR = 1.3

+2D

+2D

DAR = 1.6

+1D+110

+1D+110

+220 +110

+1D G

G

G +220

0

C

+1D G

D DAR = 1.9

+2D

+1D+110

+2D

G

+1D+110

0

E

DAR = 1.9

G

F DAR = 0.1

0

DAR = 0.1

0

G

G +1D

+1D

Figure 2: Deconvoluted mass spectra for the Fab’2 fragment of anti-Her2LCK149C conjugates reacted with compounds (A) 2 (PDS-DM3), (B) 7 (PDSmonomethyl MMAE disulfide), (C) 4 (nitroPDS-DM3), (D) 9 (nitroPDSmonomethyl MMAE disulfide), (E) 5 (nitroPDS-DM4), and (F) 10 (nitroPDSdimethyl MMAE disulfide). “0”, “1D” and “2D” peaks correspond to Fab’2 with 0, 1 or 2 payloads, respectively. “G” indicates glycosylation.


Page 7 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


respectively.24 The MMAE linker-payloads, by contrast, contained a self-immolative linker, designed to release MMAE as a free amine.18, 25 Steric hindrance next to the disulfide for MMAE disulfide-linker-payloads was varied by incorporation of zero, one, or two methyl groups on the linker. As leaving groups, we evaluated the commonly employed 2-mercaptopyridyl (PDS) and 5-nitro-2-mercaptopyridyl (nitroPDS) groups. While these leaving groups are used interchangeably throughout the literature to generate mixed protein/small molecule disulfide bioconjugates, no reported studies have focused on a side-by-side comparison of the two. Each disulfide-linker-payload was assessed for conjugation to an anti-HER2 THIOMAB™ antibody in which lysine at Kabat position 149 in each light-chain was replaced by Cys (anti-HER2LC-K149C). We reported recently that payloads connected to LC-K149C via a disulfide were more stably bound to the antibody in circulation than those connected at other Cys sites, making the LC-K149C site particularly attractive for the development of site-specific disulfide-linked ADCs.18 Conjugates were purified by cation-exchange chromatography and analyzed by LCMS to assess extent of reaction, expressed as drug-to-antibody ratio (DAR). Assuming complete and selective reaction at both engineered Cys of the THIOMAB™ antibody, DARs in these studies have a maximum value of 2.0. While variation of cytotoxin, keeping steric hindrance and leaving group the same, did not affect DAR significantly, we observed substantial DAR differences when either leaving group or steric hindrance was varied (Figure 2 and Table 1). Conjugation with the nitroPDS leaving group was particularly effective (DAR = 1.9 for analogs 3, 4, 8, and 9), whereas analogs with the unsubstituted PDS group conjugated less well (DAR = 1.3-1.7 for analogs 1, 2, 6 and 7, Figure 2A-D). The degree of improvement offered by the nitroPDS versus PDS group (∆DAR) was larger for linker-payloads with 1 methyl group on the linker versus those with 0 methyl groups (e.g., ∆DAR = +0.6 versus +0.2 for DM3 versus DM1, respectively). Strikingly, neither the MMAE nor maytansinoid disulfide-linker-payloads with 2 methyl groups next to the disulfide and a nitroPDS leaving group conjugated efficiently to anti-HER2LC-K149C (DAR = 0.1 for 5 and 10, Figure 2E and F). The lower DARs observed for the PDS versus nitroPDS disulfides did not result from incomplete reaction of anti-HER2LC-K149C with the PDS analogs, but rather from the generation of more side-products with these analogs (Figure 2A and B). Masses of these side-products (+110 Da) were consistent with reaction of the antibody Cys residues with the sulfur atom of the PDS leaving group, thereby generating a PDSantibody adduct, instead of the sulfur on the payload side of the disulfide (see byproduct pathway, Figure 1). Indeed, reduction of the conjugates with 10 mM DTT completely removed the +110 Da modifications and, expectedly, reduced the payload-antibody and interchain disulfide bonds, suggesting the byproducts are PDS-antibody adducts (data not shown). The PDS leaving group adducts were more prevalent for linker-payloads with 1 versus 0 methyl groups next to the disulfide, accounting for the lower DAR observed for linker-payloads in the former category, in the most extreme example, conjugation of PDS-DM3 (compound 2), >50% of the



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

product generated was PDS-modified (Figure 2A). By contrast, we could not detect a peak corresponding to a leaving group-conjugated byproduct for unhindered (0 methyls) or moderately hindered (1 methyl) disulfides with the nitroPDS leaving group (+154 Da). The most hindered disulfide payloads substituted with 2 methyl groups gave, in addition to minimal conversion to the desired DAR=2 conjugate, no leaving group-conjugated byproducts, leaving predominantly unreacted Cys thiols on the antibody (Figure 2E and F). We conclude that steric hindrance next to the disulfide is a major factor limiting efficient conjugation of PDS disulfides to THIOMAB™ antibody cysteines, primarily due to the generation of undesired PDS-conjugated antibody byproducts. Surprisingly, despite long-standing and widespread use of PDS disulfides for generation of mixed-disulfide bioconjugates, the leaving group side reaction we describe above is not well documented. In one laudable and recent exception, involving conjugation of unhindered PDS disulfide-MMAF payloads (0 methyl groups next to disulfide) to a Cys mutant of Shiga toxin, the authors noted the same leaving group side reaction we describe. In that report, replacement of PDS with a 3-nitro-2-mercaptopyridyl group, versus 5-nitro-2-mercaptopyridyl group in our work, was reported to improve conjugation efficiency significantly.26 Generation of hindered disulfide-linked conjugates We turned our attention next to achieving efficient conjugation for the most hindered disulfide-linker-payloads, those with 2 methyl groups next to the disulfide. This effort was motivated by the prospects for superior stability in circulation relative to less hindered disulfides and, for the self-immolative linker, increased rate of release of MMAE (Thorpe-Ingold effect). We set a high bar for success: achieving a DAR for a hindered dimethyl disulfide conjugate comparable to the DAR of 1.9 achieved with less hindered analogs. Reverse-conjugation approach. For conjugation of DM4, we wondered first whether we could activate the antibody Cys as a nitroPDS disulfide and then react this with the thiol of DM4, effectively reversing the roles of each reactant (Pathway B, Figure 1). An analogous approach has been successful to generate unhindered heterobifunctional ADCs,16, 27 unhindered direct disulfide ADCs,27 and hindered antibody-ricin conjugates28, but has not been employed to generate hindered directdisulfide TDCs. In the approach used here, the free cysteines of anti-HER2LC-K149C were reacted with 2,2’-dithiobis(5-nitropyridine) to give, quantitatively, the nitroPDS-activated antibody, which was subsequently purified. This conjugate was then reacted with excess DM4 thiol. Unlike conjugation of nitroPDS-DM4 to anti-HER2LC-K149C, “reverse” conjugation of DM4 thiol to nitroPDS-anti-HER2LC-K149C was highly effective, giving a conjugate with a DAR of 1.9 (Figure 3A). However, with this approach we detected a partial reduction of the antibody interchain disulfides, possibly due to the excess DM4 thiol employed, although the extent of reduction was not quantified (data not shown).


Page 8 of 28

Page 9 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


A

B DAR = 1.9

+2D

DAR = 1.8

+2D

G

G +1D

+1D

C

D DAR = 1.8

+2D

DAR = 1.9

+2D

+27 G

G +1D

+1D

Figure 3. Deconvoluted mass spectra for hindered dimethyl disulfide conjugates with the anti-Her2LC-K149C THIOMAB™ antibody: (A) Reverse conjugation of DM4 to nitroPDS-activated antibody (see Fig 1, Pathway B), Conjugation of compound (B) 23, (C) 20, or (D) 21. “+1D” and “+2D” peaks correspond to Fab’2 with 1 or 2 payloads, respectively. “G” indicates glycosylation. It is surprising that simply switching the position of the leaving group from the linker-payload to the antibody Cys resulted in a much more effective conjugation. Analogously, in a previous study, the second-order rate constant for reaction of penicillamine, a hindered thiol, with a PDS-activated glutathione disulfide was >100 times faster than that for reaction of glutathione with the PDS-activated penicillamine disulfide.28 Based on these results, it seems unlikely that the antibody plays a significant role in determining the reactivity preference we observed. It is possible that the Cys thiol on anti-HER2LC-K149C is intrinsically less nucleophilic than the DM4 thiol and/or can not approach an activated disulfide at the appropriate angle of attack to generate the desired product. Leaving group approach. We could not use the “reverse-conjugation” approach employed for DM4 for conjugation of the hindered MMAE linker-payload. This is



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

because the thiol form of the self-immolative linker in the MMAE disulfide-payload rapidly cyclizes to give the thiirane and is therefore unavailable for reaction with the activated antibody Cys. Instead, we were forced to consider alternative leaving groups. Given improvements in conjugation of less hindered MMAE disulfides observed when changing from PDS to nitroPDS (see above), we first synthesized the hindered dinitroPDS analog 11 and evaluated conjugation to anti-HER2LC-K149C (Table 1). However, antibody reaction with 11 gave a heterogeneous product consistent with nucleophilic aromatic substitution on the dinitropyridyl ring, but no desired MMAEconjugated product (Supplementary Figure 1). Additional 2-mercaptopyridyl-based leaving groups also failed to generate conjugates with DARs close to 2, the best of this series, compound 12, having a 4-mercapto-N-methyl-pyridinium leaving group, gave a DAR of 0.8 (Table 1). Ultimately, after exploration of a series of leaving groups unrelated to 2mercaptopyridyl (e.g., compounds 13-21, additional examples S1-S15 in Supplementary Table 1), we found that both an isothiourea and methanethiosulfonyl (MTS) leaving group enabled highly efficient conjugation of the hindered MMAE payload to anti-HER2LC-K149C, resulting in a DAR = 1.8 and 1.9 for compounds 20 and 21, respectively (Table 1 and Figure 3C-D). The isothiourea and MTS leaving groups also worked effectively for conjugation of the non-immolating DM4 payload (DAR=1.8 for both compounds 22 and 23 (Table 1 and Figure 3B), providing an alternative to the “reverse-conjugation” approach described above for generating these types of hindered conjugates. For the isothiourea-MMAE disulfide compound 20, we observed a byproduct peak that was ~27 Da larger than the DAR2 peak (Figure 3C). This peak does not match the product of reaction of the antibody Cys with the isothiourea leaving group (expected mass shift = +74 Da). It is possible that the side product results from oxidation of the antibody (+2 oxygen atoms = +32 Da). We did not observe this byproduct during conjugation of the DM4-thiourea compound 22. No reduction of interchain disulfide bonds was observed for conjugates generated from any of 10-23. While the MTS leaving group is precedented for use in constructing completely unhindered (desmethyl) disulfidelinked bioconjugates,29-30 its utility for generating more hindered conjugates suitable for in vivo applications has heretofore gone unrecognized. We could find no precedent for use of a isothiourea leaving group for construction of any disulfidelinked bioconjugate. We conclude that the MTS and isothiourea leaving groups are highly effective alternatives to PDS-based leaving groups for construction of hindered direct-disulfide TDCs. Hindered disulfide conjugation across different Cys-containing proteins. We next evaluated the generality of the isothiourea and MTS leaving groups for bioconjugation by evaluating reaction of MMAE-derived compounds 20 and 21 with three other THIOMAB™ antibody Cys mutants (anti-HER2LC-S121C, anti-HER2HC-A118C and anti-HER2LC-V205C), the Cys-containing protein human serum albumin and a Cys mutant of the cell-penetrating peptide penetratin, derived from the drosophila


Page 10 of 28

Page 11 of 28

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


antennapedia homeodomain protein (“AntP”, sequence: CRQIKIWFQNRRMKWKK). As have antibodies, both albumin and cell-penetrating peptides have been explored as potential delivery vehicles for small molecule cargos in vitro and in vivo, although disulfide-linked variants have been limited primarily to in vitro applications, likely because of low disulfide stability.2, 5, 31 As with anti-HER2LC-K149C, maximal DAR for conjugation to the additional THIOMAB™ antibody mutants was 2.0. For albumin, with a single free native Cys (Cys-34), the drug-to-protein ratio was maximally 1.0. For Cys-penetratin, we assessed conjugation efficiency as % conversion of the Table 2. Conjugation of nitroPDS hindered MMAE disulfide compound 10 and optimized analogs 20 and 21 to THIOMAB™ antibody mutants, HSA and Cys-penetratin (AntP). Values expressed are DAR for THIOMAB™ antibodies, HSA/MMAE ratio for HSA conjugates and percent conversion to AntP-MMAE for AntP conjugates. ND = not determined THIOMAB™ antibody site Compound HSA AntP LC LC HC LC (Leaving group) K149C S121C A118C V205C 10 (nitroPDS)

0.1

0.0

0.2

0.2

0.1

14%

20 (isothiourea)

1.8

1.5

1.5

1.5

0.9

90%

21 (MTS)

1.9

1.9

1.9

1.9

ND

ND

starting peptide to the MMAE-conjugated product. Hindered disulfides 20 (thiourea) and 21 (MTS) conjugated efficiently across all of the THIOMAB™ antibody sites, HSA and Cys-penetratin, especially as compared to nitroPDS compound 10 (Table 2, Figures 4-5). Compound 21 gave DAR=1.9 across all THIOMAB™ antibodies tested, whereas compound 20 gave DAR=1.8 for LC K149C and DAR=1.5 for other sites. Nitro-PDS compound 10, by contrast, gave DAR

Development of efficient chemistry to generate site-specific disulfide

Recommend Documents